This past week, I had the privilege of speaking and attending DataContent 2013 in Philadelphia and hosted by SIIA and our parent company, InfoCommerce Group.There were lots of impressive subscription-driven database companies there, from Relationship Science and PrivCo to NetProspex and Buyer’s Lab. There were also a host of non-subscription companies doing innovative things with data, from Brightscope to Circle Media.And there was lots to learn! Here’s some of the event’s best take-aways:#1. Know what technology your market is using.Christopher Grey won $1.3 million is scholarships to attend university and then created Scholly, a mobile app that helps students search available scholarships based on their own personal criteria. He’s now developing a Web-based subscription service for students. When asked by InfoCommerce founder Russell Perkins why he launched first on mobile, he replied that low-income students are actually far more likely to have access to a smartphone than to have desktop Internet access. So, sometimes, mobile first is the right strategy.#2. Follow leading indicators, not just lagging ones.This is a favorite refrain of David Chun, CEO of Equilar (check out Insider‘s Case Study on Equilar!). What he means is that while retention and renewal rates are a good indicator of the health of your subscription business, they’re “lagging” and offer little opportunity to course-correct. It’s better to looking at “leading” indicators, such as member engagement six months before renewal, so that you have a better chance of saving accounts and boosting that renewal rate pro-actively.#3. Have protocols for dealing with bad data.Bad data is a fact of life, so it’s better to have procedures and protocols to audit your data than it is to simply expect not to fail from time to time. Some companies rely on mottos like “Follow the money” since that’s harder to fake than other data points. But many also mentioned investing in human capital. As much as Big Data likes to eradicate binary operations when possible, many CEOs at DataContent continually hit upon the importance of good human controls. And Chun also recommended embedding errors in your data set before internal audits so that auditors don’t “fall asleep at the wheel.”#4. Get private data with co-op models.Many data publishers have created lucrative businesses by aggregating and cleaning up public data (often with the help of the human capital resources mentioned in tip #3). But there’s a real need for private data — from the financials of private companies to knowing the salaries of middle-managers at public companies.One way to get this private data is to create a co-op model where companies and organizations that contribute their private numbers can then access a database of other companies’ metrics. Of course, you’ll need a critical mass to start, but you may be surprised how eager some industries are to get on board with this idea.#5. The next growth market for data is in the non-English-speaking world.Despite the globalized nature of online databases, few are truly international. Relationship Science noted how they made a conscious decision to cover more of the US market than the international one because they needed to go deep before they went broad in order to sign up investors and subscribers.But the non-English-speaking world is poorly served by data tools and services. The hiccup is that rules and regulations in other countries, particularly Europe, can be more stringent. However, it’s not always imperative you set up operations abroad or index other country’s data to succeed in the international markets. Depending on your data, simply translating your database to another language could help grow your profits.